Oncogene (2003) 22, 2285–2295 & 2003 Nature Publishing Group All rights reserved 0950-9232/03 $25.00 www.nature.com/onc p16INK4a gene promoter variation and differential binding of a repressor, the ras-responsive zinc-finger transcription factor, RREB Shuling Zhang1, Xiaolan Qian2, Chanelle Redman1, Valeri Bliskovski1, Edward S Ramsay1, Douglas R Lowy2 and Beverly AMock* ,1 1Laboratory of Genetics, Center for Cancer Research, NCI, NIH, 37 Convent Drive, MSC 4255, Bldg. 37, Rm 2B-08, Bethesda, MD 20892-4255, USA; 2Laboratory of Cellular Oncology, Center for Cancer Research, NCI, NIH, 37 Convent Drive, MSC 4263, Bldg. 37, Rm 4106, Bethesda, MD 20892-4263, USA BALB/c mice are susceptible to the development of familial melanoma (Hussussian et al., 1994), and the pristane-induced plasma cell tumors, and have a rare gene is frequently inactivated in a wide variety of allelic variant in the coding region of the p16INK4a (p16) cancers by methylation or mutation (Baylin and Bestor, tumor suppressor gene that produces a protein with 2002; Jones and Baylin, 2002). Selective targeting of p16 impaired activity. We have now found that the BALB/c in mice place them at an increased risk for tumor p16 promoter has an allelic variant that may also development (Krimpenfort et al., 2001; Sharpless et al., compromise p16 activity. Following pristane treatment, 2001). Much less is known about the possible regulatory BALB/c p16 mRNA levels in B cells were lower than that defects. in DBA/2 or C.D2-Pctr1, a resistant BALB/c congenic Our recent work in knockout mice and transient strain that harbors DBA/2 chromatin surrounding the p16 transfections of p16 cDNAallelic variants has impli- locus. Four sequence variants were found between BALB/ cated the p16 gene, encoded by the Cdkn2a locus, as the c and DBA/2 in the p16 promoter region. In reporter Pctr1 plasmacytoma susceptibility/resistance gene on assays, the DBApromoter was at least four times more mouse Chr 4 (Zhang et al., 1998, 2001). Plasma cell active in driving luciferase expression than the BALB/c tumors arise in BALB/cAnPt mice following 5–12 promoter. Most of the difference in activity was localized months of chronic exposure to mineral oil, pristane, or to a single nucleotide deletion in BALB/c. This deletion silicone gels (Potter and Boyce, 1962; Potter, 1984; created a consensus binding site for RREB, a ras- Potter et al., 1994a). These tumors require IL6 (Nordan responsive transcriptional element with zinc-finger binding and Potter, 1986; Hilbert et al., 1995), which activates motifs. Transient transfections with RREB confirmed that the Ras/MAP kinase signal transduction pathway for the p16 promoter can be downregulated by RREB, in a growth (Ogata et al., 1997), and the JAK/STAT3 Ras- or Mek-dependent manner, and that the BALB/c pathway for antiapoptotic survival factors (Puthier promoter is more sensitive than DBA/2 to regulation by et al., 1999). In contrast, DBA/2N mice are solidly RREB. BALB/c mice have both regulatory and coding resistant to tumor induction with pristane (Morse III region defects that may contribute to the impairment of et al., 1980). Backcross and congenic strain analyses p16 gene function. comparing genotypes with phenotypes have identified Oncogene (2003) 22, 2285–2295. doi:10.1038/sj.onc.1206257 several susceptibility/resistance loci, including Pctr1 on mouse Chr 4 (Potter et al., 1994b; Mock et al., 1993, Keywords: p16 promoter; RREB; plasmacytoma; NIH 1997; Zhang et al., 1998, 2001; Zhang and Mock, 1999). 3T3 Studies of the Pctr1 locus focused on the identifica- tion of rare p16 alleles, as defined by differences in coding sequences (Zhang et al., 1998), and on the Introduction decreased specific activity of the p16 protein encoded by susceptible mice compared to that in resistant mice The p16INK4a (p16) gene is a tumor suppressor involved (Zhang and Mock, 1999; Zhang et al., 2001). However, in regulating cell cycle checkpoints. The p16 protein since regulatory polymorphisms may accompany coding specifically binds to and inhibits the cyclin-dependent polymorphisms, we have now explored the possibility kinases CDK4/6, which regulate cell cycle progression in that differences in the activity of the p16 promoter might G1 through phosphorylation of the retinoblastoma also contribute to the reduced p16 gene function in protein (pRB) (Goodrich et al., 1991; Ewen et al., BALB/cAn mice. Of the six transcription factors that 1993; Kato et al., 1993; Matsushime et al., 1994). have been shown to regulate the human p16 promoter, Defects in coding sequences have been associated with only three (Ets1, junB, and Id) (Alani et al., 2001; Ohtani et al., 2001; Passegue and Wagner, 2000) have consensus binding sites (Ets1, AP1, and E boxes, *Correspondence: BAMock; E-mail: [email protected] Received 19 August 2002; revised 15 November 2002; accepted 21 respectively) in the mouse. In the current study, we November 2002 have identified another transcription factor, Rreb, which RREB preferentially represses the BALB/c p16 promoter Shuling Zhang et al 2286 Figure 1 Mean p16 mRNAlevels in normal (white bars) and pristane-treated (black bars) splenic B cells isolated from plasmacytoma- susceptible BALB/cAn mice and resistant DBA/2 and C.D2-Pctr1 (p16) mice. Standard errors are indicated by a thin line. Mice treated with pristane received 0.5 ml of pristane 7 days prior to harvesting B cells. Units of mRNAexpression are arbitrary Figure 2 Schematic diagram of the mouse p16INK4a promoter coupled to the luciferase reporter gene. The diagram corresponds to sequences from À761 to +81. Arrows indicate the locations of primers for PCR. The four single nucleotide differences between DBA/ 2N and BALB/cAn promoter sequences are indicated (O ¼ deletion), as are the restriction endonuclease sites used in the cloning of the reporter plasmids can repress p16 promoter activity. It was found to bind which were similar to each other. Subsequent experi- preferentially to the promoter of BALB/c mice, relative ments involving Western blots with protein lysates from to that of DBA/2. splenic B cells also suggested that BALB/c p16 protein levels were lower than those of DBA/2 following pristane treatment (data not shown). Results p16 promoter contains sequence differences between Differential expression of p16 after pristane treatment in plasmacytoma susceptible BALB/c and resistant DBA/2 plasmacytoma susceptible BALB/cAn versus resistant mice DBA/2 and C.D2-Pctr1 congenic mice The above experiments suggested that the p16 promoter Quantitative real-time PCR was used to compare the region might be polymorphic between BALB/c and levels of p16 mRNAin splenic B cells from untreated DBA/2 mice. To assess this possibility, the promoter mice (basal levels) and pristane-treated mice. We have region of p16 (GenBank U47018) (Soloff et al., 1996) reported previously, by Northern blotting, that the basal was sequenced in BALB/c and DBA/2 strains of mice. p16 mRNAlevels in spleen are similar between the This analysis of the sequences that lie upstream from the sensitive BALB/c strain and the resistant DBA/2 strain p16 transcription start site (À761 to +81) revealed that (Zhang et al., 1998). Consistent with this result, similar BALB/c contained four sequence variants compared basal levels were found for p16 mRNAlevels in splenic with DBA/2 (Figure 2). The allelic variants (DBA/2N, B cells from BALB/c and C.D2-Pctr1 congenic mice, as BALB/cAn) were detected at the following positions assessed by quantitative real-time PCR (Figure 1). (Figure 2): À555 (G, C); À289 (G, deletion); À225 C.D2-Pctr1 mice, which are resistant, harbor a 1.5 cm (G, C); À32 (A, deletion). segment of DBA/2 chromatin that includes the DBA/2 allele of p16, on a BALB/c background. Following 7 The p16 promoter from DBA/2, a resistant strain of days of pristane treatment, p16 mRNAlevels in BALB/c mouse, is more active than the BALB/c promoter splenic B cells were reduced by approximately sixfold compared with untreated cells (Figure 1; compare white The differences in p16 mRNAexpression levels in B cells versus black bars). This level of p16 mRNAin the from BALB/c and DBA/2 following pristane treatment pristane-treated BALB/c mice was three times lower and the presence of sequence differences between the than that in the treated DBA/2 or C.D2-Pctr1 congenic, susceptible and resistant strains prompted us to examine Oncogene RREB preferentially represses the BALB/c p16 promoter Shuling Zhang et al 2287 Figure 3 p16INK4a promoter deletion and chimera studies. The length of each of the 50 deletions (a) or the chimeras (b) fused to the luciferase reporter gene is shown. Numbers indicate the 50 end of the promoter construct; constructs D-761 and C-761 contain full- length (À761 to +81) DBAand BALBpromoter sequences, respectively. Relative promoter activity (luciferase units/ mg protein lysate) following transfection into TEPC1165 plasmacytoma cells is shown (average of five experiments) whether the BALB/c and DBA/2 promoters might differ that contained combinations of the DBA/2-specific and in their activity. To test this possibility, the sequenced BALB/c-specific nucleotide variants (Figure 3b). BALB/c and DBA/2 promoter regions (À761 to +81) For each paired deletion mutant, the luciferase were cloned into a luciferase reporter gene construct activity in TEPC1165 was higher for the DBA/2 (plasmid D-761 for DBA/2 and C-761 for BALB/c). The promoter segment than for the corresponding BALB/c plasmids were transiently transfected into a BALB/c segment, although the magnitude of the difference plasmacytoma cell line, TEPC1165, to measure their appeared to vary (Figure 3a). In particular, the DBA/2 luciferase activity in a cellular context directly relevant mutant D-141, which contained only the DBA/2-specific to the disease under study.